Literature review
2.3 Methods to monitor wire health and prevent wire breakage
2.3.3 Online wire monitoring and control systems
There have been numerous efforts in developing online monitoring and diagnosing systems to study the behavior of the wire during machining and to predict the wire rupture phenomenon. Accordingly, many methods were suggested to maintain the process conditions at safe limits to avoid wire failure and enhance the precision efficiency. In the early eighties, Kinoshita et al. (1982) pointed out the importance of wire health, the various reasons behind wire failure, and developed a monitoring system to control the spark frequency. Following the footsteps, Dekeyser et al. (1988) monitored the machining process with a pulse
discriminating system and estimated the probability of wire rupture based on a thermal model to automate the EDM system. An online detection method of spark location was also developed by measuring the ratio of currents supplied through branched wires connected to a tool electrode by maintaining an adequate distance (Kunieda et al. (1990)). Several studies (Yan and Liao (1996), Rajurkar and Wang(1997), Yan and Chien (2007)) have focused on improving the overall performance of a WEDM machine by developing a knowledge-based monitoring and control system that is enabled to give indications regarding faulty operation and wire breakage. Almeida et al. (2021)comprehensively reviewed the servo control strategies for better WEDM operation and concluded that prevention of wire rupture using effective servo control approaches has further research possibilities. A few studies focus on the estimation and prevention of the wire rupture process by developing online monitoring systems that could detect and control the discharge rate at a safe level (Rajurkar et al. (1991), Rajurkar et al. (1994), Rajurkar et al. (1997)). They further developed an adaptive control system to correlate the cutting speed to the discharge rate by measuring the workpiece height with the aid of a non-linear discrete model. The machining rate is kept at an optimum level without risking the wire to frequent breakages. A similar attempt was later carried out to optimize and maintain stable input conditions without wire failure by evaluating the workpiece height using a back propagation neural network-based adaptive control system (Yan and Liao (2001)).The wire deflection caused due to external forces can be monitored online by using an optical sensor to achieve good corner accuracy at high cutting speeds (Beltrami et al. (1996)). An advanced multiple sensors system to detect the generated pulse type and discharge position was developed with the intention of zero defect manufacturing (Caggiano et al. (2016)).
Several studies reported that wire breakage was preceded by a sharp concentration of pulses, resulting in localized temperature rise, thus reducing the mechanical strength of the wire (Shoda et al.(1992)). The yielding of wire material and deformation in the plastic zone are other causes of wire breakage while sharp temperature gradients produced during spark discharges accelerate the failure process (Luo (1999)). A novel self-learning fuzzy controller was developed which increased the pulse-off time to eliminate unstable ignition delays and unhealthy sparks so that the pulsed frequency could be maintained at a constant safe level to prevent wire failure (Yan and Liao (1996)). The wire tension or rigidity should be increased
to reduce the bowing error and corner errors caused due to wire deflection and vibration (Lin et al. (2001)). Huang and Liao (2000) developed an artificial neural network (ANN) integrated expert system (ES) for error detection and maintenance of the system. A geometric path lifting method was also presented (Wang and Ravani (2003))which increases the interelectrode gap to prevent wire breakage during machining corners with very small radii.
Sarkar et al. (2011) established a wire lag compensation technique for cylindrical job cutting to modify the programmed path for higher precision efficiency. This method is useful for machining smaller radius jobs where a high level of accuracy is desired. A novel algorithm was developed based on the non-uniform distribution of sparks around the wire, which could identify the geometrical errors in an arced path, wire deflection and the variable spark gap around the wire tool (Abyar et al. (2019)). Based on the proposed model, the machining errors along the arced path were compensated with 84.8 % accuracy.
The development of a closed-loop wire tension control system was useful in deriving dynamic models of the wire feed and tension control device which improved the machining accuracy during WEDM (Yan and Huang (2004)). A closed-loop wire tension control system was further developed to ensure smooth transport system and constant tension value in order to maintain the geometrical accuracy of the machined components (Yan et al. (2004), Yan and Fang (2008)). The developed system further diminishes the probability of wire breakage during wire feeding. A constant wire tension control system was developed based on an improved wire feeding methodology and PID closed-loop control, which improved the product quality, geometrical precision, and control accuracy during machining (Chen et al.
(2018a)). A data acquisition system could identify a set of threshold input values related to spark energy, ignition delay time, and discharge current to prevent wire breakage and unstable machining conditions (Cabanes et al. (2008a)). A similar approach was adopted to propose an online surveillance system that could monitor and identify damaged cutting zones and readjust the machining parameters (Cabanes et al. (2008b)). Abhilash and Chakradhar (2020b) predicted and investigated the occurrence of wire breakage by developing an offline three-level classification model with the aid of an ANN model. The model was proved to predict the wire failure with 95 % accuracy.
Okada and his group made some important contributions in monitoring and diagnosing the wire performance during machining. In 2009, Okada et al. (2009) analyzed
the complex fluid flow and debris movement around the wire electrode to ensure stable machining performance without wire failure. The stagnation area with less flow velocity could be observed around the wire under any rate of flushing conditions. A novel method was proposed to measure the distribution of discharge position by analyzing the recorded images of a high-speed video camera during wire movement (Okada et al. (2010)). Figure 2.9 shows the high-speed observation system for a 50 μm diameter tungsten wire with a digital high-speed video camera having a recording speed of 8000 frames per second (fps). A uniform spark distribution was observed at high servo voltage, larger pulse-off time, and low wire velocity. The movements of a tungsten wire electrode were further investigated using a high-speed camera during the fine WEDM process to study the causes of wire breakage (Habib and Okada (2016)). The wire vibration amplitude and machined kerf width reduce with increased values of wire tension and the vibration amplitude parallel to the cutting direction was larger than that of the perpendicular one.
Figure 2.9 High-speed observation system of the wire electrode during WEDM (Okada et al.
(2010)) (Reproduced with permission from Elsevier)